6 papers
Breaking the Token Ceiling: Distilling Smaller, Stronger Byte Models
Kalyani Marathe, Artidoro Pagnoni, Tomasz Limisiewicz +4
Small models are made more capable through distillation from a larger one that shares their tokenization scheme. However, do distilled byte and token models behave similarly in ter…
Knowledge Distillation During Mid-Training Favors Reasoning over Factual Recall
Jacqueline He, Howard Yen, Shuyue Stella Li +9
Logit-based knowledge distillation (KD) is used to train smaller language models (LMs) via supervision from stronger teachers, but whether its benefits are consistent across traini…
Compute Optimal Tokenization
Tomasz Limisiewicz, Artidoro Pagnoni, Srini Iyer +6
Scaling laws enable the optimal selection of data amount and language model size, yet the impact of the data unit, the token, on this relationship remains underexplored. In this wo…
Precise Information Control in Long-Form Text Generation
Jacqueline He, Howard Yen, Margaret Li +7
A central challenge in language models (LMs) is faithfulness hallucination: the generation of information unsubstantiated by input context. To study this problem, we propose Precis…
Byte Latent Transformer: Patches Scale Better Than Tokens
Artidoro Pagnoni, Ram Pasunuru, Pedro Rodriguez +11
We introduce the Byte Latent Transformer (BLT), a new byte-level LLM architecture that, for the first time, matches tokenization-based LLM performance at scale with significant imp…
Predicting vs. Acting: A Trade-off Between World Modeling & Agent Modeling
Margaret Li, Weijia Shi, Artidoro Pagnoni +2
RLHF-aligned LMs have shown unprecedented ability on both benchmarks and long-form text generation, yet they struggle with one foundational task: next-token prediction. As RLHF mod…